A generic and extendable data management kernel for unified handling of geoscientific data sets
Hans Böcker and Peter Menzel and Christian Gerhards. ( 2021 )
in: 2021 RING Meeting, ASGA
Abstract
Preparation, analysis and visualization of different geoscientific datasets can often not be conducted in a single software tool or data representation, resulting in the necessity of data conversion in between steps. Although most of the time it is easy to convert simple data to an established structure, this approach won't always work when dealing with complex data, e.g. geological 3D models or simulation results. Additionally, implementing conversion code between each combination of data representations as well as validation and visualization for each representation leads to a very high implementation effort and makes the integration of additional representations into one system much more difficult. In this work, we present a generic interface definition for handling different data representations in one system and also a prototype kernel that implements this interface for handling the different complex geoscientific data representations in a unified way. Therefore, the main goal of this kernel is to build up such a basic representation. The prototype kernel consists of a well documented class structure implemented in Matlab, allowing for independent development and integration of new data handling and validation methods. Additionally, data validation and visualization methods can be implemented and performed much easier based on such a kernel compared to separate independent data representations. The most used data formats were analyzed and literature concerning the idea consulted, to ensure the kernel is set up broad enough to depict all common geodata while also being easy to expand. The kernel has been successfully tested with several different data sets from Skua-Gocad and VTK.
Download / Links
BibTeX Reference
@inproceedings{BÖCKER_RM2021,
abstract = { Preparation, analysis and visualization of different geoscientific datasets can often not be conducted in a single software tool or data representation, resulting in the necessity of data conversion in between steps. Although most of the time it is easy to convert simple data to an established structure, this approach won't always work when dealing with complex data, e.g. geological 3D models or simulation results. Additionally, implementing conversion code between each combination of data representations as well as validation and visualization for each representation leads to a very high implementation effort and makes the integration of additional representations into one system much more difficult. In this work, we present a generic interface definition for handling different data representations in one system and also a prototype kernel that implements this interface for handling the different complex geoscientific data representations in a unified way. Therefore, the main goal of this kernel is to build up such a basic representation. The prototype kernel consists of a well documented class structure implemented in Matlab, allowing for independent development and integration of new data handling and validation methods. Additionally, data validation and visualization methods can be implemented and performed much easier based on such a kernel compared to separate independent data representations. The most used data formats were analyzed and literature concerning the idea consulted, to ensure the kernel is set up broad enough to depict all common geodata while also being easy to expand. The kernel has been successfully tested with several different data sets from Skua-Gocad and VTK. },
author = { Böcker, Hans AND Menzel, Peter AND Gerhards, Christian },
booktitle = { 2021 RING Meeting },
publisher = { ASGA },
title = { A generic and extendable data management kernel for unified handling of geoscientific data sets },
year = { 2021 }
}